Margin of error
A statistical estimate of how much a result may differ from the true population value because the poll surveyed a sample rather than everyone. It does not capture every possible source of error, including poor question wording, coverage gaps, nonresponse, or flawed weighting.
Try it
See the margin of error shrink
A 50% result is about ±4.9 points — the true value is plausibly between 45.1% and 54.9% (95% confidence, simple random sample).
Quadrupling the sample only halves the margin.
Open the full Margin of Error LabIllustrative only. Real polls also carry error from wording, coverage, nonresponse, and weighting that this does not show.
Where this shows up in PolitiMind
Learn about polling methodologyConcept map
How margin of error connects to nearby ideas.
This guide is educational, not a substitute for a poll’s full methodology statement. Terms should be interpreted within the design and limitations of the specific research.